S1

SUPPLEMENTARYINFORMATION

Tuning the electronic structure of thiolate-protected 25-atomclusters by co-substitution with metals having differentpreferentialsites

SachilSharma,aSeijiYamazoe,b,cTasukuOno,aWataruKurashige,aYoshikiNiihori,aKatsuyukiNobusada,c,d,*TatsuyaTsukuda,b,c,*andYuichiNegishia,e,*

aDepartmentofAppliedChemistry,FacultyofScience,TokyoUniversityofScience,1-3Kagurazaka,Shinjuku-ku,Tokyo162−8601,Japan.

bDepartmentofChemistry,FacultyofScience,UniversityofTokyo,7−3−1Hongo,Bunkyo−ku,Tokyo113−0033,Japan. cElementsStrategyInitiativeforCatalystsandBatteries(ESICB),KyotoUniversity,Katsura,Kyoto615−8520,Japan.dDepartmentofTheoreticalandComputationalMolecularScience,InstituteforMolecularScience,Myodaiji,Okazaki,Aichi444−8585,Japan.

ePhotocatalysisInternationalResearchCenter,TokyoUniversityofScience,2641Yamazaki,Noda,Chiba278−8510,Japan.CorrespondingAuthorE-mail: negishi@rs.kagu.tus.ac.jp(Y.N.)tsukuda@chem.s.u-tokyo.ac.jp(T.T.)nobusada@ims.ac.jp(K.N.)

1.ExperimentalProcedures1.1.Chemicals

All chemicals were commercially obtained and used without further purification. Hydrogentetrachloroauratetetrahydrate(HAuCl4•4H2O)waspurchasedfromTanakaKikinzoku.Palladium(II)sodiumchloride trihydrate (PdCl2•2NaCl•3H2O), silvernitrate (AgNO3),copper(II) chloridedihydrate (CuCl2•2H2O),tetraoctylammoniumbromide((C8H17)4NBr),sodiumtetrahydroborate(NaBH4),1-dodecanethiol(C12H25SH),triethylamine,anddichloromethane(CH2Cl2)wereobtainedfromWakoPureChemicalIndustries.Methanol,ethanol, toluene, and acetone were obtained from Kanto Kagaku. Tetrabutylammonium perchlorate((C4H9)4NClO4) and trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]malononitrile (DCTB) werepurchasedfromTokyoKasei.Deionizedwaterwitharesistivityof>18MΩcmwasusedinallexperiments.

1.2.Synthesis1.2.1Au24Pd(SC12H25)18templatecluster

This synthesis includes four steps: (1) the synthesis of a mixture of [Au24Pd(SC12H25)18]2−/−/0 and[Au25(SC12H25)18]−,1(2)theoxidationofpartofthesamplestotheneutralform,(3)theroughseparationofneutral Au24Pd(SC12H25)18 from Au25(SC12H25)18 by solvent extraction, and (4) the complete removal ofAu25(SC12H25)18 from the sample by thiol etching. Specifically, a mixture of [Au24Pd(SC12H25)18]2−/−/0 and[Au25(SC12H25)18]−werefirstlypreparedaccordingtoourpreviousreportwithslightmodification.1Themixture(~50mg)waskeptinCH2Cl2(100mL)containing(C4H9)4NClO4(342mg)atroomtemperaturefor~3.5h.Itispresumedthatundertheseconditions,mostofthe[Au24Pd(SC12H25)18]2−/−clustersareoxidizedtotheneutralform,whereasmostofthe[Au25(SC12H25)18]−onesarenot.Then,anionic[Au25(SC12H25)18]–wasremovedfromthe sample by solvent extraction using acetone. The resulting sample still contained a small amount ofAu25(SC12H25)18,asconfirmedbymatrix-assisted laserdesorption ionization (MALDI)massspectrometryofthesample.Thesample(~6mg)wasdissolvedintoluene/acetone(2:1;30mL)containingdodecanethiol(2mL)andstirredatroomtemperaturefor2days.ThisetchingprocesswasrepeateduntilAu25(SC12H25)18wascompletely removed from the sample. The obtained product was further purified using a size exclusion

Electronic Supplementary Material (ESI) for Dalton Transactions.This journal is © The Royal Society of Chemistry 2016

S2

column(WatersCo.,UltrastyragelTM)toremovetheby-productproducedduringtheetchingprocedure.Thefinalproductwashighlypure,asshownintheupperpartofFig.2(a). 1.2.2M-dodecanethiolcomplex(M=AgorCu)

Metal-SC12H25complexesweresynthesizedbyamethodsimilartothatreportedforsynthesisofametal-SC2H4Phcomplex.2First,AgNO3(770mg)orCuCl2•2H2O(607mg)wasdissolvedinasmallamountofwaterandthenethanol(5mL)wasadded.Eachmixturewasvigorouslystirreduntilthemetalsaltsfullydissolved.Separately,C12H25SH(1.09mL)wasdissolvedinamixtureofethanol(7mL)andtrimethylamine(2mL).Thissolutionwasaddedtoeachmetalsaltsolution.Aftervigorousstirringfor30min,thereactionsolutionwascentrifuged.Theresultingprecipitateswerewashedseveraltimeswithamixtureofwaterandethanol(1:1)andsubsequentlywithapureethanol. 1.2.3Au24−xAgxPd(SC12H25)18trimetalliccluster

Thisclusterwaspreparedusingareportedmetalexchangemethod.2,3PureAu24Pd(SC12H25)18(5mg)wasdissolved in toluene (5 mL) and then Ag-SC12H25 (10 mg) was added. The solution was stirred at roomtemperature. Aliquots of the reaction solution (~150 μL) were taken out at regular time intervals andcentrifugedtoremovetheby-productsandunreactedAg-SC12H25.Thesupernatantwasevaporatedtodrynessandtheresiduewaswashedwithmethanolthreetimes.TheresultingproductwascharacterizedbyMALDImassspectrometry(Fig.2(a))andUV-Visspectroscopy(Fig.4(a)).ThenumberofAgatomsincludedintheclusterincreasedwiththequantityofthereactedAg-SC12complex.Forexample,weobservedtheinclusionofAgatomsupto7whenthequantityofAg-SC12wasincreasedupto15mg. 1.2.4Au24−yCuyPd(SC12H25)18trimetalliccluster

Thisclusterwasalsopreparedusingthemetalexchangemethod.2,3Specifically,Au24Pd(SC12H25)18(4mg)wasdissolvedintoluene(4mL)andthenCu-SC12H25(4mg)wasadded.Inthissynthesis,NaBH4(2mg)wasalso added to the reaction medium, similar to the reaction between Au25(SC2H4Ph)18 and Cu-SC2H4Ph.2Aliquotsof thereactionsolution (~150μL)were takenoutafter regular time intervalsandcentrifugedtoremove the by-products and unreacted Cu-SC12H25. The supernatantwas evaporated to dryness and theresiduewaswashedwithmethanol three times.The resultingproductwas characterizedbyMALDImassspectrometry(upperspectrumofFig.S6). 1.2.5Au24−x−yAgxCuyPd(SC12H25)18tetrametalliccluster

This clusterwas prepared fromAu24−xAgxPd(SC12H25)18 and Cu-SC12H25 or Au24−yCuyPd(SC12H25)18 andAg-SC12H25using themetalexchangemethod.2,3 In the formerpreparation,Au24−xAgxPd(SC12H25)18 (2mg)wasdissolved in toluene (2 mL) and then Cu-SC12H25 (2 mg) was added. In the latter preparation,Au24−yCuyPd(SC12H25)18 (2mg)wasdissolved intoluene(2mL)andthenAg-SC12H25 (2mg)wasadded.Thesolutionswerebothstirredatroomtemperature.Aliquotsofthereactionsolutions(~150μL)weretakenoutatregulartimeintervalsandcentrifugedtoremovetheby-productsandunreactedM-SC12H25(M=AgorCu).Each supernatantwasevaporated todryness and thenwashedwithmethanol three times. The resultingproductswerecharacterizedbyMALDImassspectrometry(Fig.2(b),S5andS6)andUV-Visspectroscopy(Fig.4(b)). 1.3.Characterization MALDImassspectrawereacquiredusingaspiraltime-of-flightmassspectrometer(JEOLLtd.,JMS-S3000)withasemiconductorlaser(wavelength:349nm).DCTBwasusedasthematrix.4Thecluster-to-matrixratiowassetto1:1000.

S3

PdK-edge,AgK-edgeandCuK-edgeX-rayabsorptionfinestructure(XAFS)measurementswerecarriedoutat the BL01B1 beamline at the SPring-8 facility of the Japan Synchrotron Radiation Research Institute(proposalnos.2015A1590,2015B1308,and2016A1436).Measurementswereconductedat10K.5ASi(311)two-crystalmonochromatorwasusedas the incidentbeam.Au24Pd(SC12H25)18,Au24−xAgxPd(SC12H25)18,andAu24−x−yAgxCuyPd(SC12H25)18dilutedwithboronnitridewerepressedintoapelletandmountedonasampleholder attached to the cryostat. Subsequently, Pd, Ag and Cu K-edge XAFS spectra were recorded influorescencemodeusinganionizationchamberastheI0detectorand19solidstatedetectorsastheIdetector.TheX-rayenergywascalibratedusingPd,AgandCufoilsforPd,AgandCuK-edgedata,respectively.Dataanalysis was performed with the REX2000 Ver. 2.5.9 software package (Rigaku). The extended X-rayabsorptionfinestructure(EXAFS)dataforthePd,Ag,andCuK-edgewereanalyzedasfollows.Theχspectrawereextractedbyremovingtheatomicabsorptionbackgroundusingacubicsplineandwerenormalizedtotheedgeheight.Thek3-weightedχspectraunderwentaFouriertransform(FT)intorspaceusingakrangeof3.0−15.5Å−1forthePdK-edge,3.0−14.0Å−1fortheAgK-edge,and3.0−11.0Å−1fortheCuK-edge.ThecurvefittingrangesofPd−M(M=Auand/orAg),Ag−M(M=Auand/orPd),andAg−S,Cu−Swere1.9−3.0,1.7−3.1,1.4−2.3Å,respectively.ThephaseshiftandbackscatteringamplitudefunctionsofPd−Au,Pd−Ag,Ag−Au,Ag−S,andCu−SbondswereextractedfromPdAgAu23(SCH3)18(Pd:center,Ag:edge),AgAu24(SCH3)18(Ag:edge),andCuAu24(SCH3)18(Cu:staple)usingFEFF8(σ2=0.0036,whereσistheDebye−Wallerfactor). Diffuse reflectancespectraof thesampleswereacquiredatambient temperatureusinga JASCOV-670spectrometer.Thewavelength-dependentopticaldata,I(w),wereconvertedtoenergy-dependentdata,I(E),usingthefollowingequation,whichconservestheintegratedspectralareas: I(E)=I(w)/|∂E/∂w|∝I(w)×w2.

2.Calculations

Densityfunctionaltheory(DFT)calculationswereperformedfor[Au24Pd(SCH3)18]0,1[Au23AgPd(SCH3)18]0(Fig. 5(a)), and [Au22AgCuPd(SCH3)18]0 (Fig. 5(a) and S12). In these calculations, the experimentallysynthesized clusters were modeled by replacing dodecanethiolate with methanethiolate. Geometricoptimizations of [Au23AgPd(SCH3)18]0 and [Au22AgCuPd(SCH3)18]0 were performed starting from an initialstructureestimatebasedon[Au24Pd(SCH3)18]0(ref.1).Thisinitialstructureappearstoberelativelysymmetric.However,wedidnotassumeahighdegreeofmolecularsymmetryinourcalculationsandinsteadperformedfull geometry optimization of each cluster. The TURBOMOLE package of ab initio quantum chemistryprograms6wasusedinallcalculations.Geometricoptimizationsbasedonaquasi-Newton–Raphsonmethodwere performed at the level of Kohn-Sham (KS) DFT using the Becke three-parameter hybrid exchangefunctional with the Lee–Yang–Parr correlation functional (B3LYP).7,8 The double-ζ valence quality pluspolarizationbasisintheTURBOMOLEbasissetlibrarywasusedinthecalculations,alongwitha60-electron(28-electron) relativistic effective core potential9 for the gold (palladium or silver) atoms. Optimizedstructures of these clusters with different substitution positions, and the -S(R)-[Au-S(R)-]2 staple wereobtainedatthesameleveloftheory.Theabsorptionspectraweresimulatedusingtime-dependentKSlinearresponsetheory.10–12

S4

3.Results

TableS1.TrimetallicAu24−xAgxPd(SC12H25)18andTetrametallicAu24−x−yAgxCuyPd(SC12H25)18havingSimilarMolecularWeights.

ChemicalFormula Molecularweight(Da)aAu21Ag3Pd(SC12H25)18 8191.4Au22Cu2Pd(SC12H25)18 8191.9Au20Ag4Pd(SC12H25)18 8102.3

Au21AgCu2Pd(SC12H25)18 8102.8Au20Ag2CuPd(SC12H25)18 8058.0Au21Cu3Pd(SC12H25)18 8058.4Au19Ag5Pd(SC12H25)18 8013.2

Au20Ag2Cu2Pd(SC12H25)18 8013.7Au19Ag4CuPd(SC12H25)18 7968.9Au20AgCu3Pd(SC12H25)18 7969.3

aThesevalueswerecalculatedusing“IsotopePatternSimulator”software.

Table S2. Results of Curve Fitting Analysis of Pd K-edge EXAFS Data for Au24Pd(SC12H25)18,Au24−xAgxPd(SC12H25)18,andAu24−x−yAgxCuyPd(SC12H25)18.

Cluster Bond C.N.a,b r(Å)a D.W.a,c Rfactor(%)aAu24Pd

Pd−Au 11.0(1.0) 2.769(2) 0.0058(8) 14.2Au24−xAgxPd

Pd−Ag 1.8(1.6) 2.782(24) 0.002(16) 6.7 Pd−Au 8.5(1.7) 2.726(8) 0.005(3)Au24−x−yAgxCuyPd

Pd−Ag 1.0(3) 2.724(15) 0.005(1) 10.3 Pd−Au 10.8(1.8) 2.755(9) 0.005(12)Numbersinparenthesesareuncertainties;2.769(2)represents2.769±0.002.aThevalueswereobtainedbyfittingwithPd−AuandPd−Agbonds. bCoordinationnumber.cΔDW = {(σ−σ0)

2}, where σ and σ0 represent Debye‒Waller factors of the sample and standard value (0.06),respectively.Table S3. Results of Curve Fitting Analysis of Ag K-edge EXAFS Data for Au24−xAgxPd(SC12H25)18 andAu24−x−yAgxCuyPd(SC12H25)18.

Cluster Bond C.N.a,b r(Å)a D.W.a,c Rfactor(%)aAu24−xAgxPd

Ag−S 1.4(4) 2.487(10) 0.011(5)11.5 Ag−Pd 0.9(3) 2.786(6) 0.006(3)

Ag−Au 4.5(9) 2.850(17) 0.017(3)Au24−x−yAgxCuyPd

Ag−S 1.1(3) 2.460(10) 0.009(6)14.3 Ag−Pd 1.2(6) 2.754(14) 0.013(10)

Ag−Au 4.5(12) 2.863(13) 0.015(5)Numbersinparenthesesareuncertainties;2.485(13)represents2.485±0.013.aThevalueswereobtainedbyfittingwithAg−S,Ag−PdandAg−Aubonds. bCoordinationnumber.cΔDW = {(σ- σ0)

2}, whereσ and σ0 represent the Debye‒Waller factor of the sample and standard value (0.06),respectively.TableS4.ResultsofCurveFittingAnalysisofCuK-edgeEXAFSDataforAu24−x−yAgxCuyPd(SC12H25)18.

Cluster Bond C.N.a,b r(Å)a D.W.a,c Rfactor(%)aAu24−x−yAgxCuyPd

Cu−S 2.2(3) 2.280(4) 0.012(2) 14.8Numbersinparenthesesareuncertainties;2.251(9)represents2.251±0.009.aThevalueswereobtainedbyfittingwithCu−Sbonds. bCoordinationnumber.cΔDW={(σ-σ0)

2},whereσandσ0termsrepresentDebye‒Wallerfactorsofthesampleandstandardvalue(0.06).

S5

Fig.S1Wideregionofthenegative-ionMALDImassspectraoftheclustersbeforeandafterthereactionbetweenAu24Pd(SC12H25)18andAg-SC12H25.Thesemassspectrademonstrate thateachsampledoesnotincludeclustersofothersizesinasimilarsizeregion.

Fig.S2Assignmentsofthesub-peaksappearingjustontherightof(a)thebimetallicand(b)thetrimetallicclusters. These peaks probably originate from the C2H5 radical adduct. This type of adduct peak hassometimesbeenobservedinMALDImassspectra.13−15Inthisstudy,thesepeaksdidnotdisappearevenwhenthelaserfluencewasdecreasedtothelowestvaluethatenabledustodetecttheions.

S6

Fig.S3Negative-ionMALDImassspectraoftheproductsafterreactionofAu24Pd(SC12H25)18withAg-SC12H25fora long time.Thepeakattributed toAu20Ag4Pd(SC12H25)18wasalwaysobservedwith thehighest ionintensity in the mass spectra, meaning that Au20Ag4Pd(SC12H25)18 is the most stable cluster amongAu24−xAgxPd(SC12H25)18clustersundertheinvestigatedexperimentalconditions.

Fig.S4Resultofthestabilityexperiments.Inthisexperiments,Au24−xAgxPd(SC12H25)18clusterswereleftintoluene(1.5mL)containingC12H25SH(1.5mL)at50°C.

S7

Fig. S6 Negative-ion MALDI mass spectra of the products obtained by the reaction betweenAu24−yCuyPd(SC12H25)18andAg-SC12H25.Thepeakswithasteriskscouldincludetwospecies(TableS1).

Fig.S5Wideregionofthenegative-ionMALDImassspectraoftheclustersbeforeandafterthereactionbetweenAu24−xAgxPd(SC12H25)18andCu-SC12H25.Thesemassspectrademonstratethateachsampledoesnotincludeclustersofothersizesinasimilarsizeregion.

S8

Fig.S7Negative-ionMALDImassspectraof theproductused for theEXAFS experimentsof trimetallicAu24−xAgxPd(SC12H25)18.Becausethesamplewaspreparedinashortreactiontimetoobtaininformationonthe initial substitution site of Ag, bimetallic Au24Pd(SC12H25)18 is also observed in this sample. Theestimationofthepeakarearatiointhemassspectrasuggeststhatinthissample,91.5%ofPdatomsareincludedintrimetallicAu24−xAgxPd(SC12H25)18(x≠0).ThismeansthatthespectralfeatureinthePdK-edgeFT-EXAFSspectrumofthissample(Fig.3(a))isgovernedbyPdintrimetallicAu24−xAgxPd(SC12H25)18(x≠0).RegardingAg,alltheAgatomsinthissamplearepresentintrimetallicAu24−xAgxPd(SC12H25)18(x≠0).Thus,thespectralfeatureintheAgK-edgeFT-EXAFSspectrumofthissample(Fig.3(b))showstheAgsubstitutionsiteintrimetallicAu24−xAgxPd(SC12H25)18(x≠0).

Fig.S8Negative-ionMALDImassspectraoftheproductusedfortheEXAFSexperimentsoftetrametallicAu24−x−yAgxCuyPd(SC12H25)18.Inthissample,thetrimetallicAu24−xAgxPd(SC12H25)18andAu24−yCuyPd(SC12H25)18arealsopresent,becausethesamplewaspreparedinashortreactiontimetoobtaininformationontheinitialsubstitutionsitesofAgandCu.Thepeakswithasteriskscouldincludetwospecies(TableS1).Itwasestimatedthatinthissample,30.0%ofPdatoms,43.1%ofAgatoms,and46.5%ofCuatomsarepresentintetrametallicAu24−x−yAgxCuyPd(SC12H25)18(x≠0,y≠0)undertheassumptionthatthosepeaks includetwospecieswiththesameratio.

S9

Fig. S9 (a) Experimental Pd K-edge EXAFS oscillations for Au24Pd(SC12H25)18, Au24−xAgxPd(SC12H25)18, andAu24−x−yAgxCuyPd(SC12H25)18. (b)ExperimentalAgK-edgeEXAFSoscillationsforAu24−xAgxPd(SC12H25)18andAu24−x−yAgxCuyPd(SC12H25)18.(c)ExperimentalCuK-edgeEXAFSoscillationsforAu24−x−yAgxCuyPd(SC12H25)18.The mass distributions of Au24−xAgxPd(SC12H25)18 and Au24−x−yAgxCuyPd(SC12H25)18 used in thesemeasurementsareshowninFig.S7andS8,respectively.

S10

Fig. S10Negative-ionMALDImass spectrum of (a)Au24−xAgxPd(SC12H25)18 (bottomof Fig. 2(a)) and (b)Au24−x−yAgxCuyPd(SC12H25)18(bottomofFig.2(b))observedwithalaserfluenceslightlyhigherthanthatusedtoobservenon-destructivemassspectra.TheappearanceoffragmentsionsofAuzAgxCuyPdSw(x+y+z=19–22;w=9,10)stronglyindicatesthatPdissubstitutedattheCsiteeveninthesesamples.16

S11

Fig.S11(a)Negative-ionMALDImassspectraand(b)opticalabsorptionspectraoftheproductsobtainedby the reaction between Au24−xAgxPd(SC12H25)18 and Cu-SC12H25. Au24−xAgxPd(SC12H25)18 containing adifferentnumberofAgatomsfromthatinFig.2(b)wasusedinthisexperiment.

Fig.S12Optimizedstructurefor[Au22AgCuPd(SCH3)18]0inwhichPd,AgandCuaresubstitutedatC,EandSsites,respectively.Becausecluster1ismoreenergeticallystablethancluster2by0.072eV,theopticalabsorbancespectrumof[Au22AgCuPd(SCH3)18]0(Fig.5(b))wascalculatedforcluster1.

S12

4.References1. Y.Negishi,W.Kurashige,Y.Niihori,T.IwasaandK.Nobusada,Phys.Chem.Chem.Phys.,2010,12,6219.2. S.Wang,Y.Song,S.Jin,X.Liu,J.Zhang,Y.Pei,X.Meng,M.Chen,P.LiandM.Zhu,J.Am.Chem.Soc.,2015,

137,4018.3. S.Yang,S.Wang,S.Jin,S.Chen,H.ShengandM.Zhu,Nanoscale,2015,7,10005.4. A.Dass,A.Stevenson,G.R.Dubay,J.B.TracyandR.W.Murray,J.Am.Chem.Soc.,2008,130,594.5. S.Yamazoe,S.Takano,W.Kurashige,T.Yokoyama,K.Nitta,Y.NegishiandT.Tsukuda,Nat.Commun.,2016,

7,10414.6. TURBOMOLE,version6.3,TURBOMOLEGmbH,Karlsruhe,Germany. 7. C.Lee,W.YangandR.G.Parr,Phys.Rev.B:Condens.Matter.,1988,37,785.8. A.D.Becke,J.Chem.Phys.,1993,98,5648. 9. D.Andrae,U.Häuβermann,M.Dolg,H.StollandH.Preuβ,Theor.Chim.Acta,1990,77,123–141.10. M.E.CasidainRecentAdvancesinDensityFunctionalMethodsPartI,ed.D.P.Chong,WorldScientific,

Singapore,1995,vol.1,pp.155. 11. R.BauernschmittandR.Ahlrichs,Chem.Phys.Lett.,1996,256,454. 12. R.Bauernschmitt,M.Häser,O.TreutlerandR.Ahlrichs,Chem.Phys.Lett.,1997,264,573.13. Y.Niihori,Y.Kikuchi,A.Kato,M.MatsuzakiandY.Negishi,ACSNano,2015,9,9347–9356.14. H.Qian,E.Barry,Y.ZhuandR.Jin,ActaPhys.–Chim.Sin.,2011,27,513–519.15. E.Gottlieb,H.QianandR.Jin,Chem.–Eur.J.,2013,19,4238–4243.16. S.Sharma,W.Kurashige,K.NobusadaandY.Negishi,Nanoscale,2015,7,10606.